* Codes to replicate the regression analyses in the appendix




*********************************************
* Appendix D Table D.1 (alternative models) *
*********************************************

* Model 1 (SNTV only)
cluster2 log_personalpronoun_name_prop2 Iadjust gender age incumbent log_totwin log_distpopulation distpop65 log_distprimary if year < 1996, fcluster(candidateID) tcluster(dyID)

*Linear regression with 2D clustered SEs                Number of obs =    2508
*                                                       F(  8,  2499) =    3.15
*                                                       Prob > F      =  0.0015
*Number of clusters (candidateID) =  1301               R-squared     =  0.0089
*Number of clusters (dyID) =     387                    Root MSE      =  0.7670
*------------------------------------------------------------------------------------
*log_person~e_prop2 |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
*-------------------+----------------------------------------------------------------
*           Iadjust |   .0301939   .0128662     2.35   0.019     .0049643    .0554235
*            gender |   .1021483   .0680586     1.50   0.134    -.0313089    .2356054
*               age |  -.0006169   .0019975    -0.31   0.757    -.0045338    .0032999
*         incumbent |  -.1140495   .0472177    -2.42   0.016    -.2066393   -.0214596
*        log_totwin |   .0383497   .0358069     1.07   0.284    -.0318646    .1085639
*log_distpopulation |   -.035909     .05211    -0.69   0.491    -.1380922    .0662742
*         distpop65 |  -1.329755   .5739868    -2.32   0.021    -2.455293    -.204216
*   log_distprimary |   .0184338   .0124624     1.48   0.139    -.0060038    .0428714
*             _cons |  -4.257333   .7586136    -5.61   0.000    -5.744909   -2.769757
*------------------------------------------------------------------------------------


* Model 2 (SNTV only)
cluster2 log_partyname_prop2 Iadjust gender age incumbent log_totwin log_distpopulation distpop65 log_distprimary if year < 1996, fcluster(candidateID) tcluster(dyID)

*Linear regression with 2D clustered SEs                Number of obs =    2508
*                                                       F(  8,  2499) =   47.97
*                                                       Prob > F      =  0.0000
*Number of clusters (candidateID) =  1301               R-squared     =  0.1320
*Number of clusters (dyID) =     387                    Root MSE      =  1.0841
*------------------------------------------------------------------------------------
*log_partyname_pr~2 |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
*-------------------+----------------------------------------------------------------
*           Iadjust |  -.2037261   .0307846    -6.62   0.000     -.264092   -.1433603
*            gender |   .5817517   .1066375     5.46   0.000     .3726448    .7908585
*               age |   .0134063   .0030331     4.42   0.000     .0074587    .0193538
*         incumbent |   .0031988   .0628344     0.05   0.959    -.1200141    .1264117
*        log_totwin |  -.2945392   .0522529    -5.64   0.000    -.3970027   -.1920757
*log_distpopulation |   .2304618   .0743281     3.10   0.002     .0847108    .3762129
*         distpop65 |    4.75133   .7944811     5.98   0.000     3.193421    6.309239
*   log_distprimary |  -.0326191   .0220291    -1.48   0.139    -.0758163    .0105781
*             _cons |  -9.693251   1.063797    -9.11   0.000    -11.77927   -7.607237
*------------------------------------------------------------------------------------


* create election year dummies
qui tab year, gen(year_)


* Model 3 (election fixed effects)
cluster2 log_personalpronoun_name_prop2 Iadjust gender age incumbent log_totwin log_distpopulation distpop65 log_distprimary year_*, fcluster(candidateID) tcluster(dyID)

*Linear regression with 2D clustered SEs                Number of obs =    7424
*                                                       F( 15,  7408) =   38.55
*                                                       Prob > F      =  0.0000
*Number of clusters (candidateID) =  3271               R-squared     =  0.0646
*Number of clusters (dyID) =    1869                    Root MSE      =  0.8944
*------------------------------------------------------------------------------------
*log_person~e_prop2 |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
*-------------------+----------------------------------------------------------------
*           Iadjust |   .0295894   .0122223     2.42   0.016     .0056302    .0535485
*            gender |  -.0035548   .0384138    -0.09   0.926    -.0788569    .0717472
*               age |  -.0019809   .0012403    -1.60   0.110    -.0044123    .0004504
*         incumbent |  -.0908188   .0351108    -2.59   0.010     -.159646   -.0219917
*        log_totwin |   .0820289   .0254891     3.22   0.001     .0320631    .1319947
*log_distpopulation |   .0243483   .0361324     0.67   0.500    -.0464814     .095178
*         distpop65 |   .6989273   .3725503     1.88   0.061    -.0313772    1.429232
*   log_distprimary |  -.0359159   .0098753    -3.64   0.000    -.0552743   -.0165575
*            year_1 |   .5595473    .056612     9.88   0.000     .4485716     .670523
*            year_2 |   .5021336   .0546127     9.19   0.000     .3950771      .60919
*            year_3 |   .4397056   .0527232     8.34   0.000     .3363531     .543058
*            year_4 |   .1449537   .0431924     3.36   0.001     .0602844     .229623
*            year_5 |   .0944915   .0410855     2.30   0.021     .0139523    .1750308
*            year_6 |          0  (omitted)
*            year_7 |  -.0100636   .0449822    -0.22   0.823    -.0982415    .0781144
*            year_8 |  -.1433671   .0547115    -2.62   0.009    -.2506172   -.0361169
*             _cons |  -6.033897    .484153   -12.46   0.000    -6.982974   -5.084819
*------------------------------------------------------------------------------------


* Model 4 (election fixed effects)
cluster2 log_partyname_prop2 Iadjust gender age incumbent log_totwin log_distpopulation distpop65 log_distprimary year_*, fcluster(candidateID) tcluster(dyID)

*Linear regression with 2D clustered SEs                Number of obs =    7424
*                                                       F( 15,  7408) =   92.92
*                                                       Prob > F      =  0.0000
*Number of clusters (candidateID) =  3271               R-squared     =  0.1508
*Number of clusters (dyID) =    1869                    Root MSE      =  1.1332
*------------------------------------------------------------------------------------
*log_partyname_pr~2 |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
*-------------------+----------------------------------------------------------------
*           Iadjust |  -.1808986   .0261589    -6.92   0.000    -.2321775   -.1296197
*            gender |   .3078086   .0587237     5.24   0.000     .1926935    .4229237
*               age |   .0188315   .0016803    11.21   0.000     .0155377    .0221253
*         incumbent |  -.1238406   .0384357    -3.22   0.001    -.1991855   -.0484957
*        log_totwin |  -.4774967   .0311576   -15.33   0.000    -.5385745   -.4164188
*log_distpopulation |   .0031892   .0458119     0.07   0.945    -.0866152    .0929936
*         distpop65 |  -.0904229   .4261528    -0.21   0.832    -.9258036    .7449578
*   log_distprimary |  -.0216169   .0130966    -1.65   0.099    -.0472899    .0040561
*            year_1 |   .1771958    .070482     2.51   0.012     .0390309    .3153606
*            year_2 |    .269286   .0677657     3.97   0.000     .1364458    .4021261
*            year_3 |   .4916151   .0667938     7.36   0.000     .3606803      .62255
*            year_4 |    .066491   .0419441     1.59   0.113    -.0157314    .1487134
*            year_5 |  -.2512146   .0422796    -5.94   0.000    -.3340948   -.1683345
*            year_6 |          0  (omitted)
*            year_7 |   .1261811   .0443065     2.85   0.004     .0393277    .2130345
*            year_8 |  -.2927595   .0593905    -4.93   0.000    -.4091817   -.1763373
*             _cons |  -6.230393   .6074321   -10.26   0.000    -7.421133   -5.039653
*------------------------------------------------------------------------------------


* Model 5 (candidate random effects)
mixed log_personalpronoun_name_prop2 log_Iadjust gender age incumbent log_totwin log_distpopulation distpop65 log_distprimary || candidateID:, vce(cluster candidateID) /* using Iadjust leads to the same results */

*Mixed-effects regression                        Number of obs     =      7,424
*Group variable: candidateID                     Number of groups  =      3,271
*
*                                                Obs per group:
*                                                              min =          1
*                                                              avg =        2.3
*                                                              max =          8
*
*                                                Wald chi2(8)      =     314.37
*Log pseudolikelihood = -9681.9278               Prob > chi2       =     0.0000
*
*                                          (Std. Err. adjusted for 3,271 clusters in candidateID)
*------------------------------------------------------------------------------------------------
*                               |               Robust
*log_personalpronoun_name_prop2 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
*-------------------------------+----------------------------------------------------------------
*                   log_Iadjust |   .2054529   .0329202     6.24   0.000     .1409305    .2699753
*                        gender |  -.0324967   .0386191    -0.84   0.400    -.1081887    .0431954
*                           age |  -.0025258   .0012602    -2.00   0.045    -.0049958   -.0000559
*                     incumbent |  -.1078003   .0349185    -3.09   0.002    -.1762393   -.0393613
*                    log_totwin |   .0836969   .0252593     3.31   0.001     .0341896    .1332042
*            log_distpopulation |   .2308239   .0277386     8.32   0.000     .1764573    .2851906
*                     distpop65 |  -1.015804    .251308    -4.04   0.000    -1.508359   -.5232495
*               log_distprimary |   .0226332   .0086649     2.61   0.009     .0056503    .0396162
*                         _cons |  -8.036525   .3916377   -20.52   0.000    -8.804121   -7.268929
*------------------------------------------------------------------------------------------------
*
*------------------------------------------------------------------------------
*                             |               Robust           
*  Random-effects Parameters  |   Estimate   Std. Err.     [95% Conf. Interval]
*-----------------------------+------------------------------------------------
*candidateID: Identity        |
*                  var(_cons) |   .1376483   .0125841      .1150673    .1646605
*-----------------------------+------------------------------------------------
*               var(Residual) |   .6797938   .0149997      .6510216    .7098376
*------------------------------------------------------------------------------


* Model 6 (candidate random effects)
mixed log_partyname_prop2 log_Iadjust gender age incumbent log_totwin log_distpopulation distpop65 log_distprimary || candidateID:, vce(cluster candidateID) /* using Iadjust leads to the same results */

*Mixed-effects regression                        Number of obs     =      7,424
*Group variable: candidateID                     Number of groups  =      3,271
*
*                                                Obs per group:
*                                                              min =          1
*                                                              avg =        2.3
*                                                              max =          8
*
*                                                Wald chi2(8)      =     543.54
*Log pseudolikelihood = -11053.306               Prob > chi2       =     0.0000
*
*                               (Std. Err. adjusted for 3,271 clusters in candidateID)
*-------------------------------------------------------------------------------------
*                    |               Robust
*log_partyname_prop2 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
*--------------------+----------------------------------------------------------------
*        log_Iadjust |  -.3219692   .0411338    -7.83   0.000    -.4025901   -.2413484
*             gender |   .2678645   .0591414     4.53   0.000     .1519494    .3837796
*                age |   .0183129   .0017286    10.59   0.000     .0149249    .0217009
*          incumbent |   .0472311   .0379024     1.25   0.213    -.0270562    .1215185
*         log_totwin |  -.4705176   .0312251   -15.07   0.000    -.5317177   -.4093176
* log_distpopulation |   .2021934   .0315061     6.42   0.000     .1404425    .2639443
*          distpop65 |  -.1570422   .2866024    -0.55   0.584    -.7187727    .4046882
*    log_distprimary |   .0149291   .0119171     1.25   0.210    -.0084281    .0382863
*              _cons |  -8.657569   .4486767   -19.30   0.000    -9.536959   -7.778179
*-------------------------------------------------------------------------------------
*
*------------------------------------------------------------------------------
*                             |               Robust           
*  Random-effects Parameters  |   Estimate   Std. Err.     [95% Conf. Interval]
*-----------------------------+------------------------------------------------
*candidateID: Identity        |
*                  var(_cons) |   .6566288   .0270229      .6057446    .7117873
*-----------------------------+------------------------------------------------
*               var(Residual) |   .7381175   .0183468      .7030204    .7749668
*------------------------------------------------------------------------------


* set panel data
xtset pid year


* Model 7 (candidate fixed effects)
xtreg log_personalpronoun_name_prop2 Iadjust age incumbent log_totwin log_distpopulation distpop65 log_distprimary, fe cluster(candidateID)

*Fixed-effects (within) regression               Number of obs      =      7424
*Group variable: pid                             Number of groups   =      3271
*
*R-sq:  within  = 0.0487                         Obs per group: min =         1
*       between = 0.0147                                        avg =       2.3
*       overall = 0.0032                                        max =         8
*
*                                                F(7,3270)          =     22.82
*corr(u_i, Xb)  = -0.5848                        Prob > F           =    0.0000
*
*                              (Std. Err. adjusted for 3,271 clusters in candidateID)
*------------------------------------------------------------------------------------
*                   |               Robust
*log_person~e_prop2 |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
*-------------------+----------------------------------------------------------------
*           Iadjust |   .0086113   .0147367     0.58   0.559    -.0202828    .0375055
*               age |  -.0270587    .006639    -4.08   0.000    -.0400757   -.0140416
*         incumbent |  -.0723043    .038593    -1.87   0.061    -.1479732    .0033645
*        log_totwin |  -.2054132   .0782445    -2.63   0.009    -.3588265   -.0519999
*log_distpopulation |   .0793892   .0463558     1.71   0.087    -.0115001    .1702785
*         distpop65 |   1.707971   .5245999     3.26   0.001      .679393    2.736548
*   log_distprimary |  -.0630283     .03415    -1.85   0.065    -.1299859    .0039292
*             _cons |  -5.305539   .7229885    -7.34   0.000    -6.723095   -3.887982
*-------------------+----------------------------------------------------------------
*           sigma_u |  .94908926
*           sigma_e |  .81156907
*               rho |  .57763363   (fraction of variance due to u_i)
*------------------------------------------------------------------------------------


* Model 8 (candidate fixed effects)
xtreg log_partyname_prop2 Iadjust age incumbent log_totwin log_distpopulation distpop65 log_distprimary, fe cluster(candidateID)

*Fixed-effects (within) regression               Number of obs      =      7424
*Group variable: pid                             Number of groups   =      3271
*
*R-sq:  within  = 0.0153                         Obs per group: min =         1
*       between = 0.0848                                        avg =       2.3
*       overall = 0.0604                                        max =         8
*
*                                                F(7,3270)          =      8.55
*corr(u_i, Xb)  = -0.4766                        Prob > F           =    0.0000
*
*                              (Std. Err. adjusted for 3,271 clusters in candidateID)
*------------------------------------------------------------------------------------
*                   |               Robust
*log_partyname_pr~2 |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
*-------------------+----------------------------------------------------------------
*           Iadjust |   -.048502   .0196294    -2.47   0.014    -.0869892   -.0100148
*               age |  -.0099698   .0062461    -1.60   0.111    -.0222165    .0022769
*         incumbent |   .0172954   .0405216     0.43   0.670    -.0621548    .0967457
*        log_totwin |   .2883238   .0774167     3.72   0.000     .1365336     .440114
*log_distpopulation |   .2496446   .0421014     5.93   0.000     .1670968    .3321925
*         distpop65 |   .3914634   .5405402     0.72   0.469    -.6683682    1.451295
*   log_distprimary |  -.0020386   .0371612    -0.05   0.956    -.0749002    .0708231
*             _cons |  -8.647288   .6612934   -13.08   0.000     -9.94388   -7.350697
*-------------------+----------------------------------------------------------------
*           sigma_u |  1.2752444
*           sigma_e |  .83052053
*               rho |  .70217593   (fraction of variance due to u_i)
*------------------------------------------------------------------------------------




*******************************************
* Appendix D Table D.2 (omit or impute 0) *
*******************************************

* Model 1 (omit 0)
cluster2 log_personalpronoun_name_prop2 Iadjust gender age incumbent log_totwin log_distpopulation distpop65 log_distprimary if averaging != 1, fcluster(candidateID) tcluster(dyID)

*Linear regression with 2D clustered SEs                Number of obs =    7359
*                                                       F(  8,  7350) =   43.79
*                                                       Prob > F      =  0.0000
*Number of clusters (candidateID) =  3264               R-squared     =  0.0404
*Number of clusters (dyID) =    1869                    Root MSE      =  0.9062
*------------------------------------------------------------------------------------
*log_person~e_prop2 |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
*-------------------+----------------------------------------------------------------
*           Iadjust |   .0694074   .0156671     4.43   0.000     .0386954    .1001193
*            gender |   -.027184   .0390322    -0.70   0.486    -.1036982    .0493303
*               age |  -.0018161    .001257    -1.44   0.149    -.0042801    .0006479
*         incumbent |  -.1000295   .0358279    -2.79   0.005    -.1702625   -.0297966
*        log_totwin |   .0959501   .0259202     3.70   0.000     .0451391    .1467611
*log_distpopulation |   .2385211   .0299948     7.95   0.000     .1797227    .2973196
*         distpop65 |  -1.007776   .2753562    -3.66   0.000    -1.547553   -.4679988
*   log_distprimary |   .0194312   .0089619     2.17   0.030     .0018634     .036999
*             _cons |  -8.195101   .4256942   -19.25   0.000    -9.029584   -7.360618
*------------------------------------------------------------------------------------


* Model 2 (omit 0)
cluster2 log_partyname_prop2 Iadjust gender age incumbent log_totwin log_distpopulation distpop65 log_distprimary if averaging != 1, fcluster(candidateID) tcluster(dyID)

*Linear regression with 2D clustered SEs                Number of obs =    7359
*                                                       F(  8,  7350) =  146.04
*                                                       Prob > F      =  0.0000
*Number of clusters (candidateID) =  3264               R-squared     =  0.1291
*Number of clusters (dyID) =    1869                    Root MSE      =  1.1487
*------------------------------------------------------------------------------------
*log_partyname_pr~2 |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
*-------------------+----------------------------------------------------------------
*           Iadjust |  -.1581594   .0244416    -6.47   0.000     -.206072   -.1102468
*            gender |   .2791028   .0604475     4.62   0.000     .1606084    .3975971
*               age |   .0182951   .0017002    10.76   0.000     .0149622    .0216281
*         incumbent |  -.1249545   .0391639    -3.19   0.001     -.201727    -.048182
*        log_totwin |  -.4742935   .0314676   -15.07   0.000     -.535979    -.412608
*log_distpopulation |   .2146391   .0343862     6.24   0.000     .1472322     .282046
*         distpop65 |  -1.015585   .3046235    -3.33   0.001    -1.612735   -.4184358
*   log_distprimary |    .015771   .0115079     1.37   0.171    -.0067878    .0383299
*             _cons |  -8.632758    .486006   -17.76   0.000    -9.585469   -7.680046
*------------------------------------------------------------------------------------


* Model 3 (impute 0)
cluster2 log_personalpronoun_name_prop2 Iadjust2 gender age incumbent log_totwin log_distpopulation distpop65 log_distprimary, fcluster(candidateID) tcluster(dyID)

*Linear regression with 2D clustered SEs                Number of obs =    7424
*                                                       F(  8,  7415) =   45.41
*                                                       Prob > F      =  0.0000
*Number of clusters (candidateID) =  3271               R-squared     =  0.0417
*Number of clusters (dyID) =    1869                    Root MSE      =  0.9049
*------------------------------------------------------------------------------------
*log_person~e_prop2 |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
*-------------------+----------------------------------------------------------------
*          Iadjust2 |   .0671608   .0154423     4.35   0.000     .0368894    .0974321
*            gender |  -.0303065   .0389876    -0.78   0.437    -.1067332    .0461203
*               age |  -.0017691   .0012509    -1.41   0.157    -.0042211     .000683
*         incumbent |  -.1012053   .0355421    -2.85   0.004     -.170878   -.0315326
*        log_totwin |   .0951719    .025676     3.71   0.000     .0448396    .1455042
*log_distpopulation |   .2445628   .0296878     8.24   0.000     .1863664    .3027593
*         distpop65 |  -1.030277   .2750642    -3.75   0.000    -1.569481    -.491073
*   log_distprimary |   .0207176   .0089204     2.32   0.020     .0032311    .0382042
*             _cons |  -8.264135   .4217358   -19.60   0.000    -9.090857   -7.437413
*------------------------------------------------------------------------------------


* Model 4 (impute 0)
cluster2 log_partyname_prop2 Iadjust2 gender age incumbent log_totwin log_distpopulation distpop65 log_distprimary, fcluster(candidateID) tcluster(dyID)

*Linear regression with 2D clustered SEs                Number of obs =    7424
*                                                       F(  8,  7415) =  144.84
*                                                       Prob > F      =  0.0000
*Number of clusters (candidateID) =  3271               R-squared     =  0.1274
*Number of clusters (dyID) =    1869                    Root MSE      =  1.1482
*------------------------------------------------------------------------------------
*log_partyname_pr~2 |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
*-------------------+----------------------------------------------------------------
*          Iadjust2 |  -.1589885   .0243819    -6.52   0.000    -.2067841    -.111193
*            gender |   .2786249   .0603116     4.62   0.000      .160397    .3968527
*               age |   .0182176    .001699    10.72   0.000     .0148872     .021548
*         incumbent |  -.1238647   .0391221    -3.17   0.002    -.2005552   -.0471742
*        log_totwin |  -.4698852   .0313982   -14.97   0.000    -.5314346   -.4083357
*log_distpopulation |   .2126364   .0334268     6.36   0.000     .1471103    .2781625
*         distpop65 |  -1.014722   .3041122    -3.34   0.001    -1.610869    -.418576
*   log_distprimary |   .0170146   .0114804     1.48   0.138    -.0054902    .0395194
*             _cons |  -8.601431   .4749716   -18.11   0.000     -9.53251   -7.670352
*------------------------------------------------------------------------------------




******************************************************
* Appendix D Table D.3 (control for dual nomination) *
******************************************************

* Model 1
cluster2 log_personalpronoun_name_prop2 Iadjust gender age incumbent log_totwin dualnominate log_distpopulation distpop65 log_distprimary, fcluster(candidateID) tcluster(dyID)

*Linear regression with 2D clustered SEs                Number of obs =    7424
*                                                       F(  9,  7414) =   41.60
*                                                       Prob > F      =  0.0000
*Number of clusters (candidateID) =  3271               R-squared     =  0.0429
*Number of clusters (dyID) =    1869                    Root MSE      =  0.9044
*------------------------------------------------------------------------------------
*log_person~e_prop2 |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
*-------------------+----------------------------------------------------------------
*           Iadjust |    .062564   .0150639     4.15   0.000     .0330345    .0920935
*            gender |   -.031188   .0389315    -0.80   0.423    -.1075049    .0451289
*               age |  -.0022761    .001272    -1.79   0.074    -.0047697    .0002175
*         incumbent |  -.0986713   .0354024    -2.79   0.005    -.1680701   -.0292725
*        log_totwin |   .1061423   .0263571     4.03   0.000     .0544749    .1578097
*      dualnominate |  -.0761141   .0300244    -2.54   0.011    -.1349705   -.0172578
*log_distpopulation |   .2175397   .0314659     6.91   0.000     .1558575    .2792219
*         distpop65 |  -.8939392   .2766929    -3.23   0.001    -1.436336   -.3515426
*   log_distprimary |   .0151703   .0089709     1.69   0.091    -.0024153    .0327558
*             _cons |  -7.903676   .4450711   -17.76   0.000    -8.776142    -7.03121
*------------------------------------------------------------------------------------


* Model 2
cluster2 log_partyname_prop2 Iadjust gender age incumbent log_totwin dualnominate log_distpopulation distpop65 log_distprimary, fcluster(candidateID) tcluster(dyID)

*Linear regression with 2D clustered SEs                Number of obs =    7424
*                                                       F(  9,  7414) =  189.20
*                                                       Prob > F      =  0.0000
*Number of clusters (candidateID) =  3271               R-squared     =  0.1834
*Number of clusters (dyID) =    1869                    Root MSE      =  1.1108
*------------------------------------------------------------------------------------
*log_partyname_pr~2 |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
*-------------------+----------------------------------------------------------------
*           Iadjust |  -.2065418   .0290341    -7.11   0.000    -.2634568   -.1496268
*            gender |   .2704434   .0566772     4.77   0.000     .1593401    .3815468
*               age |   .0135631   .0016533     8.20   0.000     .0103221    .0168041
*         incumbent |  -.1001222   .0396313    -2.53   0.012    -.1778107   -.0224336
*        log_totwin |   -.367956   .0314977   -11.68   0.000    -.4297004   -.3062115
*      dualnominate |   -.707799   .0369208   -19.17   0.000    -.7801742   -.6354237
*log_distpopulation |  -.0342325   .0343734    -1.00   0.319    -.1016142    .0331492
*         distpop65 |   .2398278   .3143167     0.76   0.445    -.3763223    .8559779
*   log_distprimary |  -.0333727   .0117141    -2.85   0.004    -.0563355   -.0104098
*             _cons |  -5.301557   .4858718   -10.91   0.000    -6.254004   -4.349111
*------------------------------------------------------------------------------------




*******************************************************
* Appendix D Table D.4 (fractional probit regression) *
*******************************************************

* Model 1
fracreg probit personalpronoun_name_prop2 Iadjust gender age incumbent log_totwin log_distpopulation distpop65 log_distprimary

*Fractional probit regression                    Number of obs     =      7,424
*                                                Wald chi2(8)      =     240.57
*                                                Prob > chi2       =     0.0000
*Log pseudolikelihood = -269.32553               Pseudo R2         =     0.0020
*
*--------------------------------------------------------------------------------------------
*                           |               Robust
*personalpronoun_name_prop2 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
*---------------------------+----------------------------------------------------------------
*                   Iadjust |   .0183824   .0037876     4.85   0.000     .0109588    .0258061
*                    gender |  -.0256563   .0128239    -2.00   0.045    -.0507907    -.000522
*                       age |  -.0010669   .0003928    -2.72   0.007    -.0018367   -.0002971
*                 incumbent |  -.0404099   .0123094    -3.28   0.001    -.0645358    -.016284
*                log_totwin |   .0322276    .008064     4.00   0.000     .0164225    .0480328
*        log_distpopulation |   .0689541   .0087214     7.91   0.000     .0518604    .0860477
*                 distpop65 |  -.2740301    .088044    -3.11   0.002    -.4465932   -.1014671
*           log_distprimary |   .0028566   .0027085     1.05   0.292    -.0024519    .0081651
*                     _cons |  -3.332052   .1262276   -26.40   0.000    -3.579453    -3.08465
*--------------------------------------------------------------------------------------------


* Model 2
fracreg probit partyname_prop2 Iadjust gender age incumbent log_totwin log_distpopulation distpop65 log_distprimary

*Fractional probit regression                    Number of obs     =      7,424
*                                                Wald chi2(8)      =    1651.83
*                                                Prob > chi2       =     0.0000
*Log pseudolikelihood = -298.14336               Pseudo R2         =     0.0262
*
*------------------------------------------------------------------------------------
*                   |               Robust
*   partyname_prop2 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
*-------------------+----------------------------------------------------------------
*           Iadjust |  -.0974558   .0127893    -7.62   0.000    -.1225224   -.0723892
*            gender |   .0733667    .014688     5.00   0.000     .0445788    .1021546
*               age |   .0059606   .0005331    11.18   0.000     .0049158    .0070053
*         incumbent |  -.0711079       .017    -4.18   0.000    -.1044274   -.0377885
*        log_totwin |  -.2234496   .0117301   -19.05   0.000    -.2464402   -.2004589
*log_distpopulation |   .0556491   .0121247     4.59   0.000     .0318851    .0794131
*         distpop65 |  -.4643711   .1123201    -4.13   0.000    -.6845145   -.2442277
*   log_distprimary |   .0116704   .0036977     3.16   0.002      .004423    .0189179
*             _cons |  -3.199809   .1723076   -18.57   0.000    -3.537526   -2.862092
*------------------------------------------------------------------------------------




*********************************************
* Appendix G Table G.1 (C/M vs. C/P vs. I') *
*********************************************

* Model 1 (C/M)
cluster2 log_personalpronoun_name_prop2 C_M gender age incumbent log_totwin log_distpopulation distpop65 log_distprimary, fcluster(candidateID) tcluster(dyID)

*Linear regression with 2D clustered SEs                Number of obs =    7433
*                                                       F(  8,  7424) =   54.44
*                                                       Prob > F      =  0.0000
*Number of clusters (candidateID) =  3272               R-squared     =  0.0459
*Number of clusters (dyID) =    1870                    Root MSE      =  0.9026
*------------------------------------------------------------------------------------
*log_person~e_prop2 |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
*-------------------+----------------------------------------------------------------
*               C_M |   .3722639   .0475816     7.82   0.000     .2789903    .4655374
*            gender |   -.028485   .0389441    -0.73   0.465    -.1048266    .0478566
*               age |  -.0017959   .0012425    -1.45   0.148    -.0042317    .0006398
*         incumbent |  -.1072722    .035528    -3.02   0.003    -.1769171   -.0376273
*        log_totwin |   .0812444   .0255941     3.17   0.002     .0310728    .1314161
*log_distpopulation |    .228839    .029393     7.79   0.000     .1712203    .2864578
*         distpop65 |   -.908052   .2727381    -3.33   0.001    -1.442696    -.373408
*   log_distprimary |   .0108587   .0089296     1.22   0.224    -.0066458    .0283632
*             _cons |  -8.120449   .4164552   -19.50   0.000    -8.936819   -7.304079
*------------------------------------------------------------------------------------


* Model 2 (C/P)
cluster2 log_personalpronoun_name_prop2 C_Padjust gender age incumbent log_totwin log_distpopulation distpop65 log_distprimary, fcluster(candidateID) tcluster(dyID)

*Linear regression with 2D clustered SEs                Number of obs =    7424
*                                                       F(  8,  7415) =   50.82
*                                                       Prob > F      =  0.0000
*Number of clusters (candidateID) =  3271               R-squared     =  0.0442
*Number of clusters (dyID) =    1869                    Root MSE      =  0.9037
*------------------------------------------------------------------------------------
*log_person~e_prop2 |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
*-------------------+----------------------------------------------------------------
*         C_Padjust |   .1683521    .025396     6.63   0.000     .1185686    .2181356
*            gender |  -.0280939   .0389764    -0.72   0.471    -.1044986    .0483109
*               age |  -.0018259    .001246    -1.47   0.143    -.0042684    .0006166
*         incumbent |  -.1037992   .0354936    -2.92   0.003    -.1733768   -.0342217
*        log_totwin |   .0873198    .025678     3.40   0.001     .0369836    .1376559
*log_distpopulation |   .2281905   .0296692     7.69   0.000     .1700305    .2863506
*         distpop65 |  -.9641645   .2733888    -3.53   0.000    -1.500084   -.4282448
*   log_distprimary |   .0146229   .0089395     1.64   0.102    -.0029011     .032147
*             _cons |  -8.085568   .4200186   -19.25   0.000    -8.908924   -7.262212
*------------------------------------------------------------------------------------


* Model 3 (I')
cluster2 log_personalpronoun_name_prop2 Iadjust gender age incumbent log_totwin log_distpopulation distpop65 log_distprimary, fcluster(candidateID) tcluster(dyID)

*Linear regression with 2D clustered SEs                Number of obs =    7424
*                                                       F(  8,  7415) =   45.46
*                                                       Prob > F      =  0.0000
*Number of clusters (candidateID) =  3271               R-squared     =  0.0418
*Number of clusters (dyID) =    1869                    Root MSE      =  0.9048
*------------------------------------------------------------------------------------
*log_person~e_prop2 |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
*-------------------+----------------------------------------------------------------
*           Iadjust |    .067773    .015533     4.36   0.000     .0373239    .0982221
*            gender |  -.0302941   .0389853    -0.78   0.437    -.1067164    .0461282
*               age |  -.0017744   .0012508    -1.42   0.156    -.0042264    .0006775
*         incumbent |  -.1012012   .0355425    -2.85   0.004    -.1708746   -.0315278
*        log_totwin |   .0951439   .0256752     3.71   0.000     .0448132    .1454747
*log_distpopulation |   .2441175   .0296922     8.22   0.000     .1859124    .3023226
*         distpop65 |  -1.028813   .2750447    -3.74   0.000    -1.567978   -.4896471
*   log_distprimary |   .0205914   .0089197     2.31   0.021     .0031062    .0380766
*             _cons |  -8.258962   .4217474   -19.58   0.000    -9.085707   -7.432218
*------------------------------------------------------------------------------------


* Model 4 (C/M)
cluster2 log_partyname_prop2 C_M gender age incumbent log_totwin log_distpopulation distpop65 log_distprimary, fcluster(candidateID) tcluster(dyID)

*Linear regression with 2D clustered SEs                Number of obs =    7433
*                                                       F(  8,  7424) =  160.37
*                                                       Prob > F      =  0.0000
*Number of clusters (candidateID) =  3272               R-squared     =  0.1335
*Number of clusters (dyID) =    1870                    Root MSE      =  1.1444
*------------------------------------------------------------------------------------
*log_partyname_pr~2 |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
*-------------------+----------------------------------------------------------------
*               C_M |  -.7321295   .0608428   -12.03   0.000    -.8513986   -.6128605
*            gender |   .2763463   .0602158     4.59   0.000     .1583063    .3943863
*               age |     .01823   .0016961    10.75   0.000     .0149052    .0215548
*         incumbent |  -.1104549   .0391418    -2.82   0.005     -.187184   -.0337258
*        log_totwin |  -.4467149   .0314824   -14.19   0.000    -.5084293   -.3850005
*log_distpopulation |   .2359146   .0336963     7.00   0.000     .1698603     .301969
*         distpop65 |  -1.204423   .3051535    -3.95   0.000     -1.80261   -.6062356
*   log_distprimary |   .0334034   .0118171     2.83   0.005     .0102386    .0565683
*             _cons |  -8.804744   .4778191   -18.43   0.000    -9.741405   -7.868083
*------------------------------------------------------------------------------------


* Model 5 (C/P)
cluster2 log_partyname_prop2 C_Padjust gender age incumbent log_totwin log_distpopulation distpop65 log_distprimary, fcluster(candidateID) tcluster(dyID)

*Linear regression with 2D clustered SEs                Number of obs =    7424
*                                                       F(  8,  7415) =  153.17
*                                                       Prob > F      =  0.0000
*Number of clusters (candidateID) =  3271               R-squared     =  0.1313
*Number of clusters (dyID) =    1869                    Root MSE      =  1.1456
*------------------------------------------------------------------------------------
*log_partyname_pr~2 |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
*-------------------+----------------------------------------------------------------
*         C_Padjust |  -.3504594   .0353616    -9.91   0.000    -.4197782   -.2811407
*            gender |   .2753407   .0603459     4.56   0.000     .1570457    .3936358
*               age |    .018318   .0016977    10.79   0.000       .01499    .0216459
*         incumbent |  -.1165933   .0390442    -2.99   0.003     -.193131   -.0400555
*        log_totwin |  -.4569529   .0312809   -14.61   0.000    -.5182725   -.3956334
*log_distpopulation |   .2404558   .0339166     7.09   0.000     .1739697    .3069419
*         distpop65 |   -1.12125   .3050973    -3.68   0.000    -1.719327   -.5231723
*   log_distprimary |   .0274622   .0116861     2.35   0.019     .0045541    .0503703
*             _cons |  -8.906328    .479706   -18.57   0.000    -9.846688   -7.965968
*------------------------------------------------------------------------------------


* Model 6 (I')
cluster2 log_partyname_prop2 Iadjust gender age incumbent log_totwin log_distpopulation distpop65 log_distprimary, fcluster(candidateID) tcluster(dyID)

*Linear regression with 2D clustered SEs                Number of obs =    7424
*                                                       F(  8,  7415) =  144.69
*                                                       Prob > F      =  0.0000
*Number of clusters (candidateID) =  3271               R-squared     =  0.1273
*Number of clusters (dyID) =    1869                    Root MSE      =  1.1482
*------------------------------------------------------------------------------------
*log_partyname_pr~2 |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
*-------------------+----------------------------------------------------------------
*           Iadjust |  -.1581022   .0243381    -6.50   0.000    -.2058117   -.1103927
*            gender |   .2787556   .0603068     4.62   0.000     .1605372     .396974
*               age |    .018228   .0016993    10.73   0.000     .0148969     .021559
*         incumbent |  -.1236484   .0391173    -3.16   0.002    -.2003294   -.0469675
*        log_totwin |  -.4702319   .0313977   -14.98   0.000    -.5317803   -.4086836
*log_distpopulation |   .2129195   .0334477     6.37   0.000     .1473525    .2784864
*         distpop65 |  -1.014385   .3041458    -3.34   0.001    -1.610597   -.4181729
*   log_distprimary |   .0170389   .0114882     1.48   0.138    -.0054812    .0395589
*             _cons |  -8.605433   .4751188   -18.11   0.000      -9.5368   -7.674065
*------------------------------------------------------------------------------------
